4007
Biochemistry 1981, 20, 4007-4014
Neocarzinostatin Chromophore Binds to Deoxyribonucleic Acid by Intercalation+ Lawrence F. Povirk, Nanibhushan Dattagupta, Benjamin C. Warf, and Irving H. Goldberg*
ABSTRACT: The nonprotein chromophore of neocarzinostatin was found to share many of the characteristics of classical intercalators in its interaction with DNA. Viscosity studies with PM2 DNA indicated that the DNA helix unwinding induced by the chromophore was 0.82 times that of ethidium or 21O. Electric dichroism of the chromophore-DNA complex showed that each bound chromophore molecule lengthened DNA by 3.3 A and that absorbance transitions of the chromophore at 3 15-385 nm were oriented approximatelyparallel to DNA bases, as expected for an intercalated aromatic ring. Binding to DNA induced strong hypochromicity and a pronounced red shift in the absorbance spectrum of the chromophore. Spectrophotometrictitrations suggested at least two types of chromophore binding sites on DNA; one type of site was saturated at rb = 0.125 chromophore molecule/nucleotide, but binding to additional sites continued to at least rb = 0.3.
These physical-chemical studies were performed at pH 4-5 in order to keep the chromophore stable, but chromophore bound to an excess of DNA at pH 7 showed a stable absorbance spectrum identical with that seen at pH 4-5, suggesting that a similar type of binding occurs at neutral pH. Chromophore which had spontaneously degraded in pH 8 buffer did not bind to DNA at all, as judged by absorbance spectroscopy. The degree of protection afforded by DNA against spontaneous chromophore degradation implied a dissociation constant of approximately 5 pM for the DNA-chromophore complex at neutral pH and physiological ionic strength. Supercoiled DNA was nearly twice as effective as relaxed DNA in protecting chromophore from degradation, providing additional evidence for intercalation at neutral pH. Comparison of absorbance, fluorescence, and dichroism spectra suggests that the naphthalene ring system is the intercalating moiety.
E x t r a c t i o n of biologically active nonprotein chromophores from the protein antibiotics neocarzinostatin (NCS)' and auromomycin has greatly clarified the mechanism of action of these drugs (Napier et al., 1979; Kappen et al., 1980a,b; Ohtsuki & Ishida, 1980; Suzuki et al., 1980; Woynarowski & Beerman, 1980). The isolated NCS chromophore is even more active than the native chromophore-protein complex in degrading DNA, and a number of experiments with both isolated DNA and cultured cells support the view that the protein does not interact with DNA at all but serves mainly as a camer to protect the labile chromophore from spontaneous degradation (Kappen et al., 1980a; Kappen & Goldberg, 1980; Povirk & Goldberg, 1980). Recently we presented evidence for reversible binding of the biologically active NCS chromophore to DNA under conditions where it did not degrade the DNA, Le., in the absence of the required sulfhydryl cofactors (Povirk & Goldberg, 1980). This binding unwound superhelical pMB9 DNA, suggesting intercalation by NCS chromophore. NCS chromophore contains 2-hydroxy-5-methoxy-7-methylnaphthoate and 2,6-dideoxy-2-(methylamino)galactosecovalently linked to a highly unsaturated ClSHloO4substituent of undetermined structure (Albers-Schonberg et al., 1980). The naphthoic acid derivative is a binuclear fused ring system potentially capable of intercalation. The ClSHI0O4 substituent may also contain one or more aromatic rings. However, several compounds, including platinum complexes and covalently bound benzo[a]pyrenediol epoxide, have been reported to unwind DNA by inducing local denaturation rather than by intercalating (Cohen et al., 1979; Gamper et al., 1980). Therefore, we
undertook a more detailed characterization of the DNAchromophore complex.
t From the Department of Pharmacology, Harvard Medical School, Boston, Massachusetts 02115 (L.F.P., B.C.W., and I.H.G.), and the Department of Chemistry, Yale University, New Haven, Connecticut 0651 1 (N.D.). Received December 10, 1980. This work was supported by U.S.Public Health Service Research Grants GM 12573 and CA 15583 from the National Institutes of Health. L.F.P. was supported by fellowships from the National Science Foundation and the National Cancer Institute.
Experimental Procedures Preparation of Chromophore. Four ampules ( 8 mL) of clinical NCS were dialyzed overnight against distilled water, lyophilized, dissolved in 0.8 mL of 20 mM sodium citrate, pH 4, and again lyophilized. The white residue was resuspended in 0.8 mL of methanol and kept at 0 OC for 10 min, and the precipitate was pelleted by centrifugation. The colorless supernatant (0.68 mL) was decanted, and 0.17 mL of distilled water was added to improve its pipetting properties; the resulting solution contained NCS chromophore at a concentration of 0.20-0.26 mM and was stored at -70 OC. The precipitate, when redissolved in 4 mM sodium acetate, pH 5, retained 30% of the initial NCS absorbance at 330-350 nm, indicating that approximately 70% of the chromophore had been removed. Usually the precipitate was immediately reextracted with 8 mL of acetic acid to remove remaining chromophore and dissolved in 8 mL of 0.1 M Tris, pH 8. This solution was subjected to XAD-7 chromatographyas described by Napier et al. (1980) to obtain purified NCS apoprotein. Although some chromophore extraction procedures result in preparations with high proportions of various inactive chromophore degradation products (Kappen et al., 1980a), there were several indications that the procedure described above produced nearly pure active chromophore. Analysis of the chromophore used in our studies by high-pressure liquid chromatography (kindly performed by Dr. M. A. Napier) revealed only one major (90%) and one minor (10%) peak, which eluted in the same positions as the major and minor active chromophore components obtained in an extraction of NCS with a 1:l methanol-acetic acid mixture (Napier et al., 1981; Albers-Schonberg et al., 1980). No detectable amounts of the several inactive components found in methanol ex-
* Abbreviation used:
NCS, neocarzinostatin.
0006-2960/81/0420-4007%01.25/00 1981 American Chemical Society
4008
B I oc H E M I ST R Y
tractions of undialyzed, lyophilized clinical NCS (Kappen et al., 1980a) were seen. The presence of pH 4 citrate buffer apparently acted like acetic acid to stabilize the chromophore against degradation during extraction. The absorbance spectrum of our chromophore in methanol was indistinguishable from that of the purified major active component (Napier et al., 1981). Reconstitution of our chromophore and apoprotein preparations reproduced an absorption spectrum indistinguishable from that of native NCS and markedly different from the spectrum of free chromophore. Exposure of chromophore solutions to fluorescent lighting was avoided. DNA. Sonicated, SI nuclease treated calf thymus DNA (Povirk & Goldberg, 1980) was used for spectral studies. For dichroism studies, this same DNA was fractionated on a Sepharose 4B column as described by Hogan et al. (1978). Bacteriophage PM2 DNA was prepared as described by Espejo et al. (1969). Supercoiled pMB9 and Col E l plasmid DNAs were isolated as described by Blair et al. (1972). Relaxed, closed circular Col E l was prepared by using the Agrobacterium tumifaciens relaxing enzyme as described previously (Povirk et al., 1980). Linear pMB9 DNA was prepared by sonicating supercoiled pMB9 for 20 s with the microtip of a Branson cell disruptor (Hyde & Hearst, 1978). This DNA had no detectable supercoiled molecules remaining, but sedimented as a monodisperse peak in neutral sucrose at approximately half the s value of nicked circular pMB9, and had 3 1% hyperchromicity. Thus, it consisted almost entirely of linear duplex DNA. All DNA concentrations are expressed in moles of nucleotide per liter. Protein Determinations. For determination of extinction coefficients, 1-mL solutions of measured absorbance (A280 -0.04-0.12) were prepared, and norleucine (10 nmol) was added as an internal standard. Duplicate samples were then hydrolyzed and subjected to amino acid analysis (kindly performed by W. E. McGovern, Biophysics Research Laboratory, Harvard Medical School, and the Division of Medical Biology, Brigham and Women's Hospital, Boston). The analysis was in all cases consistent with the known composition of NCS apoprotein (Meienhofer et al., 1972), and the concentration was determined from the alanine content, assuming 17 alanine residues per molecule. The maximum variation between duplicates was 6%. DNA Unwinding. Reversible relaxation of PM2 DNA was monitored by viscometry (Waring, 1976). The flow time of a 2.5-mL solution of PM2 DNA (0.288 mM nucleotides) in 20 mM sodium citrate, pH 4, plus 10% methanol was measured in triplicate in a capillary viscometer. Serial additions of 10-40 p L of 0.6 mM ethidium bromide were made, and the point of full relaxation was determined from the point of maximum viscosity of the solution. DNA unwinding by chromophore was determined from its ability to decrease the amount of ethidium required to achieve full relaxation. The viscometer had a flow time of 90 s for water at 22 OC and 185 s for 9:1 water-methanol at 4 O C . Electric Dichroism. Orientation of 3 10-385-nm absorption transition moments of the chromophore was determined by measuring the electric linear dichroism of chromophore bound to short (150 base pairs along) rodlike DNA (Hogan et al., 1979). A solution containing 150 pM DNA and 6 pM chromophore in 5 mM sodium acetate, pH 5, plus 1.6% methanol was prepared and kept at 4 O C , except during a brief (5 min) degassing. Very nearly quantitative binding of chromophore is expected under these conditions (Povirk & Goldberg, 1980), and this was confirmed by the observation that there was no detectable spontaneous degradation of
P O V I R K ET A L .
chromophore for at least several hours, as monitored by either generation of 490-nm fluorescence (Povirk & Goldberg, 1980) or changes in the absorbance spectrum (Napier et al., 1981). This solution was placed in a modified temperature-jump apparatus at 4 OC and exposed to an electric field of 12-32 kV/cm, which oriented the DNA parallel to the field [see Hogan et al. (1978) for details]. The chromophore absorbance parallel to the field was measured before (Ao)and after (All) orientation to give the reduced dichroism p = 3/2(All - AO)/Ap The angle a between the transition moment of the chromophore and the DNA helix axis is then given by p = 3/2(3cos2 a - 1)9, where @ is a field-dependent orientation factor, 0 I 9 I1 (O'Konski et al., 1959). p was determined at several field strengths, E, and extrapolated to 1/E = 0 (Le., to perfect DNA orientation). For DNA alone at pH 5 this extrapolation gave p = -1.15 at 265 nm, very close to the values measured previously for DNA at neutral pH (Hogan et al., 1978). Because of the photolability of the chromophore, a fresh solution was used for each dichroism measurement, so that exposure to light (from a xenon lamp with monochromator) was reduced to several seconds. Longer exposures resulted in a loss of fine structure in both the absorbance and electric dichroism spectra of the chromophore-DNA complex. Changes in DNA length induced by chromophore binding were measured under similar conditions by determining the rotational diffusion coefficient 6 of the DNA in the presence and absence of chromophore. The DNA was oriented by rapid application of the electric field, and the orientation relaxation time, 7 = 1/(66), was determined by recording the absorbance of light polarized parallel to the field (Hogan et al., 1978). The equation of Broersma (1960) relating 6 to length was used to calculate length changes for the complexes. Chromophore Protection Assays. Recently we showed (Povirk & Goldberg, 1980) that binding of chromophore to DNA could be monitored by protection against spontaneous chromophore degradation. Chromophore (10 pL) in an 80% methanol-20% aqueous solution was added to 0.5 or 2.5 mL of DNA in sodium phosphate buffer which had been adjusted to an apparent pH of 7.0 with NaOH. Chromophore degradation rates were determined from kinetics of generation of 490-nm fluorescence (attributable to an inactive degradation product) with a fluorometer thermostated at 22 OC. These kinetics could be described by a single exponential whose rate constant ( R = 1/7) is proportional to the fraction of unbound chromophore (Povirk & Goldberg, 1980). DNA Strand Break Measurements. Reaction mixtures (100 pL) containing 30 pM DNA (a mixture of supercoiled 14Clabeled Col E l and enzymatically relaxed, closed circular 3H-labeled Col El), 10 mM 2-mercaptoethanol, 0.1 M NaC1, 50 mM Tris, pH 8, and 0-16 nM chromophore were prepared and incubated at 37 OC for 15 min. Chromophore (0-0.8 pM) was prepared from stock solutions by dilution in 20 mM sodium citrate at 4 OC,and 20 pL of these dilutions was added as the last component to incubation mixtures to yield the final concentrations indicated. The number of strand breaks per molecule was determined by measuring the loss of covalently closed relaxed or supercoiled molecules, using neutral sucrose gradients containing ethidium bromide (Povirk et al., 1979). Results Extinction Coefficient of the Chromophore. Although the chromophore has a distinct absorbance spectrum and its molecular weight (661) is known (Albers-Schonberg et al., 1980), the limited amount of material available made a direct weight determination impractical. Therefore, we chose to
NEOCARZINOSTATIN CHROMOPHORE AND DNA
03
-1
I
c
041
I
I
0
30
60
J 90
CHROMOPHORE ADDED (uL ) FIGURE 1: Titration of NCS apoprotein with isolated chromophore. Aliquots (10 pL) of chromophore were added to 1 mL of 20 mM sodium citrate, pH 4,containing 8.3 pM apoprotein (A2,* = 0.12) which had been purified by XAD-7 chromatography. The ultraviolet-visible spectrum was recorded, and absorbance at the indicated wavelengths was plotted after correction (> k32, enhancement of strand breakage will reflect only the change in k23, not the change in k32. Thus enhancement of breakage may
+
4014
BIOCHEMISTRY
be less than enhancement of binding. Conversely, supercoiling may increase kS,that is, increase the rate of conversion of labile intermediates into strand breaks, for example, by increasing the “breathing” rate and/or the mechanical stress in the DNA molecule. Such phenomena may explain the anomalously large enhancement of breakage seen with camptothecin as well as the enhancement seen with nonintercalators. In light of these complications, enhanced breakage of supercoiled DNA must be regarded as suggestive but not conclusive evidence for involvement of intercalative binding in DNA breakage. The findings that DNA lengthening by NCS chromophore is nearly the equivalent of one base pair and that unwinding is comparable to that of ethidium indicate that K23is significantly greater than 1; that is, most of the DNA-bound chromophore is intercalated, and intercalation makes a significant contribution to the binding energy. It is also possible that intercalation plays a more specific role in positioning the redox-active C,sH,o04substituent for attack on DNA sugars (Hatayama & Goldberg, 1980). References Albers-Schonberg, G., Dewey, P. G., Hensens, 0. D., Leisch, J. M., Napier, M. A., & Goldberg, I. H. (1980) Biochem. Biophys. Res. Commun. 95, 1351. Bauer, W., & Vinograd, J. (1968) J . Mol. Biol. 33, 141. Blair, D. G., Sherall, D. J., Clewell, D. B., & Helinski, D. R. (1972) Proc. Natl. Acad. Sci. U.S.A. 69, 2518. Braithwaite, A. W., & Baguley, B. C. (1980) Biochemistry 19, 1101. Broersma, S . (1960) J . Chem. Phys. 32, 1626. Cohen, G. L., Bauer, W. R., Barton, J. K., & Lippard, S. J. (1979) Science (Washington, D.C.) 203, 1014. Espejo, R. T., Canelo, E. S., & Sinsheimer, R. L. (1969) Proc. Natl. Acad. Sci. U.S.A. 63, 1164. Gabbay, E. J., DeStefano, R., & Baxter, C. S . (1973) Biochem. Biophys. Res. Commun. 51, 1083. Gamper, H. B., Straub, K., Calvin, M., & Bartholomew, J. C. (1980) Proc. Natl. Acad. Sci. U.S.A. 77, 2000. Hatayama. T., & Goldberg, I. H. (1980) Biochemistry 19, 5890. Hinton, D. M., & Bode, V. C. (1975) J . Biol. Chem. 250, 1061.
Hogan, M., Dattagupta, N., & Crothers, D. M. (1978) Proc. Natl, Acad. Sci. U.S.A. 75, 195.
P O V I R K ET A L .
Hogan, H., Dattagupta, N., & Crothers, D. M. (1979) Biochemistry 18, 280. Hyde, J. E., & Hearst, J. E. (1978) Biochemistry 17, 1251. Jung, A., Lewis, R. S., & Kohnlein, W. (1980) Biochim. Biophys. Acta 608, 147. Kappen, L. S . , & Goldberg, I. H. (1978) Nucleic Acids Res. 5 , 2959. Kappen, L. S . , & Goldberg, I. H. (1980) Biochemistry 19, 4786. Kappen, L. S., Napier, M. A., & Goldberg, I. H. (1980a) Proc. Natl. Acad. Sci. U.S.A. 77, 1970. Kappen, L. S., Napier, M. A., Goldberg, I. H., & Samy, T. S . A. (1980b) Biochemistry 19, 4780. Li, J. L., & Crothers, D. M. (1969) J . Mol. Biol. 39, 461. Lown, J. W., & Chen, H.-H. (1980) Biochem. Pharmacol. 29, 905. Meienhofer, J., Maeda, H., Glaser, C. B.,& Kuormizu, K. (1972) Science (Washington, D.C.) 178, 875. Mong, S . , Huang, C. H., Prestayko, A. W., & Crooke, S . T. (1980) Cancer Res. 40, 3313. Napier, M. A., Holmquist, B., Strydom, D. J., & Goldberg, I. H. (1979) Biochem. Biophys. Res. Commun. 89, 635. Napier, M. A., Kappen, L. S . , & Goldberg, I. H. (1980) Biochemistry 19, 1767. Napier, M. A., Holmquist, B., Strydom, D. J., & Goldberg, I. H. (1981) Biochemistry (in press). Ohtsuki, K., & Ishida, N. (1980) J . Antibiot. 33, 744. O’Konski, C. T., Yoshiro, K., & Orttung, W. H. (1959) J . Phys. Chem. 63, 1558. Povirk, L. F., & Goldberg, I. H. (1980) Biochemistry 19,4773. Povirk, L. F., Hogan, M., & Dattagupta, N. (1979) Biochemistry 18, 96. Povirk, L. F., Hogan, M., Dattagupta, N., & Buechner, M. (1981) Biochemistry 20, 665. Suzuki, H., Miura, K., Kumada, Y., Takeuchi, T., & Tanaka, N. (1980) Biochem. Biophys. Res. Commun. 94, 255. Wang, J. C. (1974a) J . Mol. Biol. 87, 797. Wang, J. C. (1974b) J . Mol. Biol. 89, 783. Waring, M. (1970) J . Mol. Biol. 54, 247. Waring, M. J. (1976) Eur. J . Cancer 12, 995. White, H. L., White, J. R., & Cavallito, C. J. (1971) Prog. Mol. Subcell. Biol. 2, 262. Woynarowski, J. M., & Beerman, T. A. (1980) Biochem. Biophys. Res. Commun. 94, 769.
